Why a Traditional CDP Couldn’t Handle Our Subscription Data at FabFitFun

To the outside world, subscription seems like the dream business model: a user signs up for your service, and then they pay you consistently month over month. For anyone who’s worked for a subscription company, you know it isn’t quite that easy. Acquisition, retention, and personalization are the keys to longevity and the reason why powerhouses like Netflix and Amazon have succeeded.

At FabFitFun, one of our biggest challenges is personalizing offers to our customers and reactivating churned users. For several years we had been relying on a traditional Customer Data Platform (CDP) to power our marketing use cases, but the platform simply wasn’t flexible enough to support our use cases.

We’ve done a lot of work to personalize offers for our customers and reactivate churned users. We had been struggling to support many of our use cases in our previous CDP—and we got much better results by switching to a Composable CDP. In this blog, I’ll dive into why we made this shift and what it helped us accomplish.

A Quick TL;DR: What is FabFitFun?

Before I get into the nitty-gritty details, I think it’s important to share some quick context so you can get an inside scoop on the challenges we faced at FabFitFun.

• Our Business: FabFitFun is one of the largest lifestyle subscription membership platforms. We curate products from 400+ unique brands and deliver quarterly boxes to our members.
• Our Business Model: Our revenue is generated in two forms: (1) subscriptions to our quarterly box drops and (2) exclusive member sales and recurring product “reFill” subscriptions.
• My Role: I work directly with our data and marketing team to help ensure we deliver the best member-facing experiences and personalization–everything from the products in our boxes to the exclusive sales and offers on our website and the overall shopping experience.
• Our Use Cases: Here at FabFitFun, we’re constantly focused on three things: optimizing our ad spend across paid media, improving our member experience using personalization, and reducing customer churn.

Our Data Was Too Complex

The problem is that we have complex data, and traditional CDPs aren’t really built to handle anything outside of behavioral events. What do I mean by complex data? Specifically, objects that are solely unique to our business.

In our case, we don’t have a typical checkout flow. We have active members and expired members, and both are engaging with our e-commerce site. Our members can add items to their cart and swap them in or out for a certain period before their box ships. Our systems had additional complexities like reviews, verified purchases, and the fact that members could swap their box for store credit, which they could redeem later.

Our previous CDP had a very limited data model based on users and events. We could use it to understand how people were interacting on our site, but if we wanted to truly understand what our best-selling products were, that information only lived in our data warehouse. We knew we were just getting a single slice of pie when we could have been having the whole pie.

The problem with bad tooling is that it creates a culture where you only do things if you have to. These limitations meant that there were really important use cases that we were handling manually or simply not solving at all. In particular, there were three key problems:

• Our paid media team had no way of differentiating between members and non-members, so they were wasting spend targeting the wrong users.
• Our support team had to hop between tools to address tickets because their data was scattered across tools.
• Our marketing team could only target select cohorts because they didn’t have the data they needed to personalize to all of our members.

We knew the data to solve these issues was in our data warehouse, which ultimately led us to the Composable CDP.

Why We Chose Hightouch

Hightouch enabled us to use our data warehouse to power all of our marketing use cases. It was flexible enough that we didn’t have to think about how our data was structured or modeled; we could just use it to drive value immediately.

With Hightouch, we can build any audience and sync any data point to our operational tools immediately without waiting on our data team. It’s essentially like our marketing team has access to the warehouse without actually having access to the warehouse. This means we’ve been able to solve use cases that were previously impossible.

For example, one of our biggest challenges was that we had no ability to differentiate between active and expired members, and we knew we could solve this by leveraging the first-party data in our warehouse to create one global exclusion list. With Hightouch, we’re able to sync this data to any of our ad platforms to ensure we’re not targeting active members.

Lifecycle marketing was another big use case for us, and with our previous solution, we could only personalize based on user-completed events. We couldn’t leverage custom data models, historical data, traits, or specific attributes about our members. With Hightouch, we’re able to sync customer profiles directly to Braze. Anytime the underlying data in Snowflake changes, we can push that data directly to Braze to ensure our marketing team is working off the most up-to-date information.

Another critical use that Hightouch is helping us solve is customer support. We have hundreds of agents, and every quarter, when we release a new box, support tickets skyrocket. Before Hightouch, our agents would have to copy the name/email of the support ticket into our various to understand the context of that ticket. Now, we can sync all of our customer data directly to our support tool so they can operate out of one system and access all of the information they need immediately, which greatly lowers the average handling time of each ticket.

Takeaways Since Adopting Hightouch

There’s a lot of misconception in the MarTech world right now that Composable CDP solutions are really hard to implement, and I can tell you that this just isn’t true. We deployed Hightouch on Redshift initially and then migrated to Snowflake, and Hightouch worked seamlessly across both.

Hightouch is just the new normal for us. I can’t imagine going back to the way things were before. We can make weekly recommendations before, after, or during our sales, which has doubled shopping participation across new members. You can learn more about this in the full case study, but here are the key takeaways we unlocked:

• 8% increase in sales from personalized emails
• 2x increase in e-commerce purchases for new members
• 1-minute reduction in average handling time for all support tickets.

To be honest, we’ve just scratched the surface with Hightouch. We have big plans to continue leveraging our first-party data to accelerate growth, unlock new channels, and optimize existing ones.